Hilde Kühne
Also known as Hilde Kuhne
Kuehne coauthored the CVPR 2025 study CAV-MAE Sync, which concerns fine-grained alignment in contrastive audio-visual masked autoencoders. Kuehne coauthored LeGrad, published at ICCV 2025. The shared study develops an explanation method for vision transformers using feature-formation sensitivity. VOLD, at CVPR 2026, is a shared study on transferring reasoning from language models to vision-language models through on-policy distillation. Her work also includes test-time adaptation for video reasoning.
Timeline
-
2025
Audio and visual alignment
Kuehne coauthored the CVPR 2025 study CAV-MAE Sync, which concerns fine-grained alignment in contrastive audio-visual masked autoencoders.
-
2025
Explaining vision transformers
Kuehne coauthored LeGrad, published at ICCV 2025. The shared study develops an explanation method for vision transformers using feature-formation sensitivity.
-
2026
Reasoning across model types
VOLD, at CVPR 2026, is a shared study on transferring reasoning from language models to vision-language models through on-policy distillation. It also lists work on test-time adaptation for video reasoning.
Sources and credits
Photo: ellis.eu; Hilde Kühne profile photograph; creator not stated. License: Not stated by source. Source: ellis.eu. Cropped to a square and resized. Ask us to remove or replace this photo.